open eye and closed eye in android by Android eye detection and tracking with OpenCV

限于喜欢 提交于 2019-12-04 16:36:01

I've no idea whether there is any library for that, but using technique descirbed in article Eye-blink detection system for human–computer interaction by Aleksandra Królak and Paweł Strumiłło (you can download it here and here and here is some simplified version) in my opinion is a good option. Generally this technique is quite simple:

  1. Find eye (or both eyes). Remember this part of image as a template.
  2. In next frame use some kind of corellation (authors used normalized cross correlation method, but you can try with other types - OpenCV has few types of correlation methods implemented) to find region similar to your template. Place with highest correlation value (most likely) will be eye.
  3. If correlation value > some_threshold_value than eye is open, otherwise it is closed.

Threshold value used by authors is in the article, but when i was using this technique i've used other value so most likely you will need to find value for your case on your own.

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!